Escolar Documentos
Profissional Documentos
Cultura Documentos
Summary
For various cross-sections the hydrodynamic coefficients of two-dimensional cylinders are determined by forced oscillation tests
and by theoretical computations. The.purposeof thisstudy is to check the theoretical basis of the computations for all three possible
modes of motion and to establish the influence of section shape in this respect.
Theresults show good agreement for heaving and forswaying, while there is-a fair correspondence for rolling. Apart from deviatitins
due to experimental inaccuracies appreciable differences between theory and experiment only exist for the coefficients- of those terms
hich dissipate energy in sway and roll There viscous effects are distinctly present especially for sharply edged sections in roll
The wave exciting terms in the two-dimensional case are measured and compared to calculations as well. So a complete-set of hydro-
dynamic quantities for the coupled motions of cylinders in beam waves is presented.
a priori that for these motions the neglect of viscous MeLeod iitl Ilsich on the other lind found tlit the
and non-linear effects is just as permissible as for the wave damping only accounted for 20 to 50 per cent. of
symmetric heaving. Possibly eddy formation plays a the total damping.
more important role. For both reasons a theoretical In heaving some experiments vcre performed in the
and experimental investigation of the basic infinitely thirties by Dirnpker [7] and Holstein [8]. They imi-
long cylinder is ofgreat valuc.Cylinders with 5 different tated two-dimensional conditions for a circle, a wedge
crosssections have been oscillated in sway and roll and a nd a recta agIe i n a sinn Il ta n k . The forner author
the measured hydrodynamic coefficients are compared only investigated free oscillations. He determined the
with those, computed by potential theory. Coupling damping decrement, while the increase in natural
terms of sway into roll and vice versa are included. period with respect to the period calculated allowed
To complete the picture also the wave exciting forces him to give an indication of the added mass Holstein.
and moment on the restrained cylinders have been on the other hand, also perforiiied forced oscillations
obtained by measurement. They are compared with and measured the progressing waves. His results may
theoretical results as well. So a complete set of hydro- still be useful, although they arc not very accurate and
dynamic quantities is presented, by which the two- are possibly influenced by wave reflection.
dimensional case, that is the motion pattern of infinitely Further it is interesting to note that Dimpker found
long cylinders in beam waves, can be analysed. a departure from pure two-dimensional conditions in
forced heaving. For a combination of heave amplitude
and frequency, exceeding a certain limit, a standing
2 Historical development
wave system along the length of the cylinder developed.
The subject rolling has an important place in the According to his investigations this is a pure hydro-
literature since 1860. Attention concentrated especially dynamic phenomenon, not depending upon section
upon roll damping by determining extinction curves shape, surface tension or accidental circumstances
for free floating models and even for actual ships. The during the test.
value of severaFof these tests may be questioned when In 1949 Urscll [9] had also indicated a general way
they were performed in small basins or at full scale in to come to a theoretical solution of (lie boundary value
docks or harbours. Besides the influence of the induced problem in two dimensions. Grim [IO, Il] and Tasai
swaying and of the position of the centre of gravity i [12, 13] extended this principle from the circular to
a certain condition of loading upon the results of the elliptic cylinders and Lewis-forms, while Porter [14]
tests does not seem to have been recognized fully. In ultimately formulated the solution for heaving of an
1933 Serat [3] used small cylindrical models for extinc- arbitrarily shaped cylinder. Now in principle the way
tion experiments to study the effect of different forms was free to investigate the influence of form, of fre-
and of the position of the centre of gravity on roll quency of motion, and of the coupling effects between
damping more fundamentally. Although he used cylin- sway and roll in detail. But first the validity of the
ders in principle he did not imitate two-dimensional theoretical approach had to be established by experi-
conditions. But in 1937 Baumann [4] did for a large. ment. Naturally this was first tried for the most simple
circular cylinder. He recognized that added mass, case of heaving. The experimental difficulties are very
moment of inertia and damping in roll were zero for great, however, and it is not surprising that only a few
this section, a fact which allowed him to determine the experiments of the actual two-dimensional case are
added mass and damping in sway for his freely floating known. Tasai [15] measured the wave heights produced
model. He only investigated one frequency of motion, by forced heaving cylinders. Porter [14] measured the
but his results are of a remarkably correct order. Ursell total vertical force on a heaving circular cylinder and
(1949) [5] calculted the outgoing waves for a forced the pressure in a number of points along the contour.
rolling motion at very low frequencies by potential Paulling and Richardson [16] carried out the most
theory. He found that for a well-rounded rectangle of extensive experiments so far. For fourdifferent sections
BIT 2.52 roll damping would vanish. This theoret- the vertical force and the pressure in 4 to 6 locations
ical result vas experimentally verified by McLeod and was recorded, both its magnitude and phase. Wave
1-Isiek [6]. For the first time now the roll axis was fixed heights were measured as well.
in space and situated in the water surfice. Their experi- The results of these experiments vere such that the
ments were not fully convincing, but they agreed fairly theoretical prediction was contirmedsubstantially. Thus
well with the predicted results, despite the fact that the for heaving only some details remain to be investigated,
tests were carried out at the natural frequency of the as discussed in the introduction.
cylinders, which was not so low that the condition Now time has certainly come to direct the attention
w -e O of Ursell could simply be considered satisfied. again to rolling and.swaying. The situation is quite dif-
253
fcrcnt with respect to these motions. Theoretical predic- a51 = hydrodynamic mass or mass moment of inertia
tions for some Lewis-frms have been presented by in the. i-mode of motion,
Tasai [:13], but as far as the author knows not a single a,j = mass coupling coefficient in the i-equation by
experimental check in these fields is available. And it motion in thej-mode,
has already been stated in the introduction that the b5 = damping coefficient against motion in the i-
validity of the theory cannot automatically be extended mode,
to these cases. lt is especially questioned whether the b = damping coupling coefficient in the i-equation
flow condition when rolling can be described ade- by motion in the j-mode,
quately by potential theory. Cil hydrostatic restoring coefficient against a dis-
placement in the i-direction,
As experimental procedure two methods are possible. Cli = hydrostatic coupling coefficient in the i-equation
When the cylinder has aforced oscillation in one mode by a displacement in the j-direction,
of motion, either the pressure along the cylinder con- Y = horizontal wave force (Y) when freely floating
tour or the force required to sustain the motion can in waves or external force (Y0,0) when forcedly
be measured. Pressure measurement has the advantage oscillated in still water,
that it allows the most direct and most detailed com- z = ditto in the vertical direction,
parison with theory. On the other hand itrequires high. K = ditto, moment about O.
accuracy, complicated equipment and extensive anal- The coefficients c1, and c53 can by definition be deter-
ysis mined by pure hydrostatics.
The force measurement involves the prcssre integra- By simple reasoning the equations (3.1) can be
tion over the body surface, so only the overall result simplified greatly. The horizontal displacement is not
can be compared with theoretical predictions. opposed by any restoring force, so e = C:i = C#, = O.
In DeIft experience has been gained with the latter The vertical motion is symmetric with respect to the
way of testing and an advanced measuring technique z-axis and. canot produce any lateral forces or mo-
has been developed [li]. Therefore the force measure- ments; therefore = b, = c, = a,s = b#: C = O.
ment as to amplitude and phase has been accepted to A static heel j does not generate a horizontal force,
obtain the experimental results. or c, = O. But due tO differences in the immersed and
emerged wedge when heeling about a fixed axis in
3 The mathematical model for motions in two space, in general C:# O. Then the mathematical
dimensions model is reduced to
'nethod [9]. The generalization of this method to the type measured the outgoing waves at a distance of
waying and rolling of arbitrarily shaped cylinders has 10 m from the centre line of the model.
been given by De Jong [18] in Delit. It was program- The cylinders occupied the whole width of the tank;
med in Algol-60 by the Computer Department of the the clearance between the end bulkheads and the tank
Technological University at Deift, while the calcul- walls being only some millimeters. The cylinders were
tions vere carried out on their TR-4 computer. constructed of wood with transverse and longitudinal
stiffening. The principal data are summarized in table 1.
The cross-sections are shown in figure 1. The trans-
5 The experiments
formation coefficients (rom the unit circle are given
5.1 The general set-up in the appendix.
The experiments were done in the main basin of the Measurements were done for three amplitudes of mo-
Delit Shipbuilding Laboratory. Its dimensions are tion over a frequency range ranging from w = 1 rad/sec
142 m length and 4.20 m width. As the draughts of the to 12 rad/sec. For heave and sway the amplitudes were
cylinders and the set-up for the various modes of mo- 0.01, 0.02 and 0.03 m; for roll 0.05, 0.10 and 0.20
(ion were different the water depth varied between radians (2.86, 5.73 and 11.46 deg.).
1.80 m and 2.25 ni. About halfway its length a very
stiff bridge was constructed across the tank. At this 5.2 The heave experiments
bridge the motion mechanism was built up. At both All of the seven sections were tested in heaving. The
nds of the tank beaches damped out most of the sinusoidal motion was directly produced by a vertical
generated waves. A wave height meter of the resistance oscillator.
By a method of harmonic analysis the in-phase and
quadrature component of the first harmonic part of
the force signal were obtained. This measuring tech-
nique has been described in [2] and [17]. The outgoing
waves were directly recorded on a UV-recorder and
read manually.
From equation (3.6) it is easily derived that the heave
coefficients are obtained by the following relations
and the ultimate results were combined to obtain th however, to obtain the required quantities for rolling
required forces and the moment about the point of about the intersection of centreline and waterline. The
intersection of the centreline and the waterline; Just equations are presented below. Just as in swaying it
as in heaing the wave height signal was written on a was not accepted a priori that vertical forces would be
U V-recorder. The two Lewis-forms were not inves- absent. The two Lewis-forms were not tested in
tigated in sway. rolling either.
The equations (14), (3.5) and (3.6) show that the The coefficients can be obtained with th aid of
required coefficients can be derived from the measure- equation (3.4), (3.5) and (16) as follows
ments as follows
K'jq'j cos CK - (c + 0Gng.)0
Yoc 3COS Cy
W
2
--
(1%). - ,fl
W'.0
+ OGs,-0Ga-0G2
'j sin C)
b, =
b=+ SiflC
K0
a. - COS C
+0G'b'0Gby#_0G2b;.y
(5.2)
yosc acos ev
K00 sinc, a, = 0Ga, (5.3)
b, = +
W).0
0Gb, (1)2
,
method failed, howcvcr. At that moment it was not the error in a for the middle and higher frequencies
ssible to switch ovcr to a rccordcr and thereforc no is still fairly large. The supposed added moment of
..ieasured wave heights are presented for (h is case. inertia is negative and this must probably be attributed
to experimental errors or to a systematic mistake in the
5.5 The !;:easzrcInc',,i of tt'a ve forces determination of the body inertia. The latter was deter-
The wave forces and (lie wave moment were recorded mined dynamically by allowing the cylinder to roll
on a UV-recorder. The recordings were analysed freely in air when attached to a spring and measuring
ma nu a l'y. the period. Before doing so the spring-apparatus was
The incoming waves were masurcd 30 ni in front calibrated by known weights. If this method should
of the models This distance could not be made shorter turn out not to be accurate enough (in this case in-
as the reflection against the model of the starting waves dicating too large inertia) the a for all sections should
inflUenced the recording so soon that for the longer be too small. Indeed this is the general picture. It can-
waves no time was left to obtain a stationary picture not be decided conclusively, however, although the
The wave moment lias to be corrected for the rolling test in air (see below) produces another indica-
distance 0G as well. The right hand side of equation tion in this respect.
(3.5), being the wave moment about G, is measured as The coupling coefficient a for the rolling circle
a whole. So that for the case of the BIT = 4 and 8 deviates rather much from zero as well. But this was
due to a weak point in the connection of the cylinder
rectangle sections the measured K is to be compared
to (lie theoretical vale axis to the bridge across the tank. The construction
hinged slightly, thus producing a small sway notion
(K = K%.+0GYW (5.5) of the cylinder. The horizontal force measured is fully
accounted for by the mass inertia force maw2, when
}. and Z,,, are always measured directly. a is a horizontal displacement of 0.5 x l0 m for
lt has to be noted that (5.5) is not an algebraic but 4a = 0.05 and of 2x iO m for /'a = 0.20. As the
a'vector-equation. Due account has to be taken of the circle was the first cylinder tested in rolling this point
phase relations of the respective quantities. was detected soon and eliminated by considerable
strengthening of the weak element. So for the other
56 Discussion of i/le experi!nenla! accuracy sections this error inay will be much smaller.
In a complex test as this it is hardly possible to analyse
the separate sources of possible errors and to estimate By these results it will also be clear that the coupling
their magnitude. it is possible, however, to obtain an coefficients of the rectangle with BIT = 4 will be very
idea of the overall accuracy by the results of some inaccurate because they have the same order of mag-
special tests. nitude as the measurements in the circle case.
lii the first place the circular cylinder was rolled.
The output of all dynamometers should be zero now, The rolling test in air was done with the rectangle
except for the inertia of the body itself. In the second BIT = 4. Only the moment to sustain the oscillation
place some oscillation experiments were exactly was measured. The body inertia measured in this test
cpeated in air. This was obtained by lowering the was indeed some four per cent lower than that being
water level so that nothing was changed. It will be determined With the spring apparatus. This is of the
understood that in this way only sway and roll cxperi same order as a for the circle rolling in water;
ments arc possible, for the buoyancy is indispensable figure 6.1. lt supports the idea that the determination
to balance the weight for vertical force measurements. of I is not very accurate, as suggested above, and that
The only non-zero result now should be the mass or thereby the measured a, will be too small for all sec-
mass moment of inertia of the cylinder itself. tions. Of course it is impracticable to determine the
inertia of all sections by an oscillation test in air, so
The rolling circle in water produced little output that this discrepancy has to be accepted.
indeed. The results are presented in figures 6.1 and 6.2. The measured b4, in air was of the same order as
The total error in dampingcoefflcient b and coupling that of (lie circle; see figure 6.1.
coefficient b is very small; compare the results for
the other sections in figures 6.3 through 6.10. In the In swaying the test in air was performed with the
low frequency range the mass coeflicients a and a, rectangle BIT = 8. The inertia force measured, com-
show distinctly that they are in error. A very small pared to the ideal inertia force lflYaW2, was slightly
absolute error in the measured force or moment causes smaller for low frequencies and slightly higher for high
large deviations in the coefficients in this range. But frequencies. The deviation at the low frequency side
258
A \! 2g 2g
AB
turns out to be completely within the absolute;mea- scatter and the consistency of the experiments is very
suring accuracy, but at the high frequency side it is satisfactory. Only in the low frequency 'range'
meaningful (about + IO per cent). (coj(B/2g) < 0.50) deviations appear, especially in
The in-phase moment about the centre of gravity the added mass. This is due to' experimental inaccura-
was less satisfactry, especially for frequencies above cies. No non-linear effects with the amplitude of oscil-
co = 7 sec'. This suggests that the centre of gravity lation could be detected, not even for the triangle.
of the ballasted model was not exactly in the required The agreement with the theoretical computations is
position and/or that the torsional rigidity of the also good. The lines drawn represent calculations ac-
cylinder itself allowed for deformations and unpropor- cording tothe best section fit, the dotted lines.according
tional loading of the gauges. This must make the a#, to the corresponding Lewis form. In general the experi-
(at least for the rectangle with BIT = 8) rather in- mental values arc very slightly higher at the high
accurate. frequency side. With this in mind 'all coefficients for
all sections are systematically more closely predicted
By the latter experiment in air an estimate could be
with the aid of the actual section fit than by a Lewis
made of the overall error in the various sway coeffi-
form. This can best be demonstrated with figures 3.1,
cients. lt is presented in table 11.
3.2 and 3.3. The sections form a three parameter
Of course, strictly speaking, these numbers onlyapply
family; the transformation coefficients are given in the
to the rectangle BIT = 8. But it does not seem too
crude an assumptin to suppose that the order of 'appendix in table A-I. Within the two-parameter Lewis
form they are identical. The 'dotted lines in figures 3.2
magnitude is also valid for the other sections. By doing
and 3.3 do correspond with the drawn ines in fig. 3.1.
so the results of the sway experiments, presented in
The differences, shown by the computations when all
figures 4.1 'through 4.8, are very satisfactory. The.
the coefficients are taken into' account, are fully con-
measured ay and for the circle for instance (which
firmed by the experiments.
should, be zero just, as a, and b,) are practically all
The differences between the two theoretical calcula-
within this error range. Since in general little scatter
tions are generally not of much importance for the
was experienced and' the measurements could be repro-
duced nicely as well the experimental accuracy has 'no
sections tested. This proves that for heaving the
breadth-draught ratio and area coefficient are the most
doubt been' considerable. In the opinion of the 'author
important parameters; at least for the cross-sections
the results approach the limits of experimental possibi-
considered.
lities for complicated tests of this type.
62 For st'aying
6 Discussion of the results
The added mass and damping in sway, together with
6.1 For/zeaving the two coupling coefficients into roll are presented in
The non-dimensional added mass and damping coef- figures 4.1 through 4.8. The measurement of vertical
ficients in heaving are presented in figures 3.1 through forces to determine the coupling coefficients a, and
3.5. b, of sway into heave showed a large scatter and the
The experimental points have directly been derived results all fell within the experimental accuracy. There-
from the measurements without smoothing a force fore no significance can be attached to them and a:,
curve through the measured points. There is little and b, 'should be taken equal to zero.
259
The rneasurng points showcd itL1c scatter and no 63 For the generated waves
ystematic discrepancy between the various amplitudes The amplitudes of the outgoing waves in swaying and
of oscillation, except for the triangle (figures 4.3 and heaving are shown in figures 5.1 through 5.7. As said
4.4). A smooth force curve has been drawn through in section 5.4 no measurements of the waves in rolling
the measured points, and this curve was further used are prcsent
for the determination of the cocrncients. Therefore the The same,points as raised in the discussion of swaying
measured coefficients do determine a continuous curve and heaving apply here. The wave measurements are
and only one symbol is used in the graphs. This way consistent with the force measurements. The only ex-
of working was preferable because the roll coupling - ception is figure 5.2, where the measured waves seem
coefficients are rather sensitive and it is logical to to be closer to the Lewis prediction while the damping
smooth the experimental data at the origin. coefficient obtained from force measurement is closer
In swaying experimental errors in the low-frequency to the actual prediction. An explanation cannot be
range are hardly present. Only the mass coefficients given. Probably it must be attributed to an error in
show some deviations below w/(B/2g) = 0.25. The the calibration of the wave height meter in this test.
coupling coefficients into roll for the circle (figure 4.2) The scatter at the high frequencies in heaving with the
should be zero, while for the rectangle BIT = 4 (figure rectangle BIT = 8 (figure 5.7) is caused by the fact that
4.7) they are very small and will be very unreliable. the cylinder starts slamming at the water surface here.
These facts have already been discussed in the sections It is noteworthy that the scale for tjy0 in the upper
.6 and 5.3. half of the figures is twice that for C/; in the lower
In general experiments and computations correspond half. Swaying generates higher waves than heaving
very satisfactorily. The agreement in added mass and above a certain frequency. Below that frequency hardly
mass coupling coefficients is good. The b, and b, do any waves and damping appear in horizontal motion.
show the influence of viscosity, however. In this respect
the results for the three rectangles and for the triangle 6.4 For rolling
arc convincing. In ligure 4.5 the theoretical and experi- The coefficients, derived from the rolling tests are
mental bry coincide for BIT = 8 and get more oIT for presented in figures 6.1 through 6.10. As in swaying
BIT = 4 and 2 at the higher frequencies. It can be the vertical force measurements did not lead to a value
understood that the flow pattern about the deeper of a1 and different from zero.
immersed sections suffers more from separation than Again as in swaying a smooth force curve has been
about the shallow one. Difference with sway amplitude drawn through the measuring points and further data
cannot yet be noted. lt culminates in a still stronger have been taken from this curve. If necessary three
eddy formation at the sharp edge of the triangle, which curves for the various amplitudes of motion were used.
also increases with sway amplitude (ligure 4.3). The The experimental results for the circle have been dis-
b, shows the same tendencies. From the above it may cussed in section 5.6.
tentatively be concluded that apparently the generated The rectangle BIT = 2 shows satisfactory measure-
eddy can roughly be considered as an additional ments a, and b, are in full agreement with the
phenomenon which hardLy disturbs the pressure dis- computations. In b the effect of viscosity can be
tribution over the section given by potential theory noted, while a suffers from errors which has been
and thus predicts the added mass and mass coupling discussed in section 5.6 as well. The data fit the trend,
coefficients correctly. however. The triangle (figures 6.3 and 6.4) shows
Just as in heaving the experiments show a slight exactly the same picture, apart from the fact that the
tendency to produce somewhat higher values for added influence of viscosity is much more pronounced, as
mass and damping than the theory at the high frequen- can be expected. Of course, the relative importance of
cies. Probably this is a small systematic error at the the viscous effects in rolling will be much larger than
high frequencies of motion, where high demands are in swaying, because the wave damping is of a smaller
imposed on the structural set-up. order. There is a large difference between the various
Also in swaying the actual section fit improves the amplitudes of motion in and The differences
theoretical prediction. Generally the difference between increase approximately linearly with 4, but the starting
the two computations is not great, however. The point is not necessarily the theoretical line. These facts
actual section contour obtains somewhat more im- suggest that the roll coefficients b and b, can be
portance for a correct prediction of the coupling represented by
coefficients.
b,,(w,40) = {b#$w)}heor + ib,,(w) + c1(o) &
The two sections forming a three parameter family (6.1)
with the circle were not tested in sway. by4,(W,lJa) = {byqs((0)}theor + E.b(w) + c2(w) .
260
The value of the /'s and oF c and c2 dillers from scc the body in no way disturbs the incident wave system.
lion to section. For sccLons not cdgcd too sharp!y Since the potential of this wave system is known the
apparently only e1 difFers appreciably from zero. This pressure distribution over the body surface is also
is the case e.g. with the rectangle BIT = 2. known and' the forces required can be obtained by
rS in Svaying the eddy formation does not seem to integration. lt is clear that this view can only provide
dtstrov ilic pressure distribution over the sectiOn, since a rather rough approximation of the actual situation.
and arc satislictoriIy predicted by potcntitI In fact the whole diffraction problem of the waves
theory. about the body has to be solved before a realistic
: COI1t)IriSOfl between the actual section fit and pressure distribution can be obtained. This is a very
thc Lewis form is not 'cII possible, since the experi- difficult task and an approximate solution has been
mental results are not accurate enotigh in this respect. sought by applying correction terms to the forces
The rciiii ni ng t'o sect ions : the recta nglcs with obtained by t he FroudcK rylov hypothesis for the
B,T = 4 and S present difficulties. Since the centre of relative motion between the body and the water
rotation in these cases was not situated ir. the waterline particles. lt has been shown several times that these
the coefficients had to be corrected for the distance 0G corrections are quite large and that this procedure
as described in section 5.4. The results for a and leads to reasonable results for the exciting force and
are shown in figures 6.7 and 6.9 It is clear that the moment in head waves.
influence of the position of G on the coefficients is Haskind found relations between the exciting forces
very large. In the damping coefficients a large contri- and the far field velocity potential for forced oscillation
bution of viscosity is noted but they fit the trends fully. in calm water. Thus an actual solution of the diffrac-
That the effect of viscosity is more pronounced than tion problem can be avoided and the exciting forces
for the rectangle BIT = 2 is understandable because are simply related to the damping coefficients. Newman
rolling takes place about a point well above the water- [19] elaborated this further and found for two-dimen-
line. No explanation can be given for the fact that in sional sections in beam waves the following expressions
figure 6.7 /a = 0.05 gives higher experimental values for
b than = 0.10 or0.20, while infgurc 6.9 the mutual Ya=a[ui_b,y]}
order is as might be expected. The measurements of a i
for both sect ions are very low and evidently greatly
= (6.2)
unreliable. A direct explanation is not available, but
reasons may b found in the sections 5.4and 5.6.
The computed values of 2,j/B/Ja for thc rectangl = c[-2 b]i
and th triangle aie presented in figure 7; the circle
does not generate any waves. The upper half of this Non-dimensionahized by the respective hydrostatic
figure shows a strong dependence on DIT-ratio. Ursell force and moment the results of these calculations are
[5], derived analytically for the very low frequencies shown in figures 9.1 through 10.5. The computations
that a well rounded rectangle of BIT = 2.52 would not have been performed with the damping coefficients fdr
generate waves in rolling and consequently have no dam- the actual section fit.
ping. To check this some additional computations were On the other hand Motora [20] used the method of
made for the samesection at various draughts. They are applying correction terms to the FroudcKrylov force.
given in figure 8. The following range of B/T-ratios Following his procedure the dotted lines in the graphs
was covered: 1-2-2.50-3.20-4-8 and 16, while five are obtained. Again the nmss and damping coefficients
frequencies of motion were selected: w ..j(B/2g) = for the actual section fit have been used. No way is
0.50-0.75-1.00-1.25 and 1.50. Vanishing roll damping available for a computation of the moment according
is found for all frequencies somewhere between BIT = to these lines. How important the correction terms are
2.50 and 4.00. The actual point is dependent on fre- is shown in figure 9.1 for the circle. The calculation by
quency. Below this point the roll damping of deep and the FroudcK rylov hypothesis (including Smith-effect)
narrow sections increases sharply at low frequencies. is far off, especially in the horizontal force. When even
Above this point the shallow and broad sections offer the Smith-effect in the incoming wave is not taken into
great advantage at the high frequencies of motion. account the calculation has become purely static and
a line at the point 1.00 over the whole frequency range
6.5 For rite wave forces and i/ic ii'arc Inomeilt will result. The mass correction term doubles the hori-
The theoretical computation of exciting forces on fixed zontal force in the very long waves for the circle and
bodies is generally based on the so-called Froude- even triples it for the triangle (figure 9.5).
Krylov hypothesis, which states that the presence of In general the computations according to Newman
261
obtain importance for the coupling cthcts of sway The inlluencc of B/T-ritio is large. Especially roll
and roll. Since it is hard io judge about tite mag- damping nearly vanishes in the importa ni p'-
flitLide of these coupling effects beforehand it is tical B/7range. The added moment of inertia is
recommeded that a better approximation of the rather depending on BIT as well.
section contour than a Lewis-form is used fr IO. The iiiathematical model, presented in 'section 3,
theoretical predictions. When a computer pro- will probably be very useful for actual computa-
gram is present it requires little extra effort to tions of ship motions in beam waves. All of the
do so. coefficients can be obtained by computation ac-
5. The wave exciting forces are well predicted by both cording to potential theory when the section con-
theoretical methods. The Newman-method, which tour is known, 'but four of them will have to be
includes the diffraction of the incoming waves, is corrected for viscous effects: h, b. h and h,,.
preferred because it isfundamentally more straight- The correction to the first one is no doubt the most
forward and much simpler. lt also gives a correct important, while that to the last one will be negli-
theoretical prediction of the wave moment. A dis- gible in several cases; the two coupling coefficients
advantage is that it does not present phase angles. can be taken equal.
6,. The calculation of the horizontal force and mo- Il. The coupling effects of sway into roll and vice
ment according to the FroudeKrylov hypothesis versa are mixed up With the influence of the verti-
and the lg wave approximation largely under- cal position of the. centre of gravity.. Therefore
estimates the actual wave force and nioment. they may be very different for various cases.
The Smith-effect causes an important decrease in
the exciting forces and moment, even for the fairly 8 Acknowledgement
long waves.
For those who are familiar with experiments in ship
Since viscous effects will mainly be due to separa- hydrodynamics it will be clear that this programme
tion and eddy formation and not to skin friction could not have been carried' out without the assistance
it is believed that scale effects are. of minor im- of many. There is hardly any member of the personnel
portance for oscillation tests to determine the of the Shipbuilding Laboratory who has not attributed
coefficients, or for ship motion tests.. At the size to it in some way or another.
of the models, used in thc.experiments, also surface The author wishes to express his sincere gratitude to
tension will only have a negligible influence. all of them.
Shipbuilding Laboratory. Technological University Dclft, NIWMAN, J. N., The exciting forces on fixed bodiesin waves.
November 1963. Journal of Ship Research, vo!.. 6, no. 3, December 1962,
l8. J0NG, B. DE, Bcrekcning van dc llydrodynanlischc cofl- p. 10-17.
cienten van oscillcrcndc cylinders. Report 174 of the Ship- MOTORA, S., Stripwisc calculation of hydrodynamic force.s
building Laboratory, Technological University Delft, March dueto beam waves Journal of Ship Research, vo!. 8, no. 1,
1967.. June 1964, p. 1-9.
APPENDIX
The tanslormation coefficients for all of the seven sections arc presented in the tables below.
Table A-I. Three-parameter family Table A-III. Other sections as actual section fit
BIT 2 4 8 1.155
fi 09992 0.9983 9966 0.50
a1 O +0.292531 +0.547939 -0.314152
a, -0.139447 -0.122406 -0.086768 +0.172370
264
ii
o
PA
0:11 .0
' Y2g
Q i. 0.01 m
o 002m
o 003m
bfW
pAY 29
_
i:
0.
:.i.J
P PII
Q 02 050 075 100 25 150
-wYi-
Fig. 3.1. Added mass and darnpingcocfficient in heaving
-
S 200
265
t::
t
o
075 1,00 125 o
1.50 1.75 200
Trw.
92g
25 050 075 100 125 1.50 1.75 200
125
O z,.0.OIm 2
011.2 1. 8
O O.02m
O 003m
100
pAT 2g
pAT 2g
1h
p,
.5
t 4
o
0.
2.
2.
\V
Q"
pA pA
i':: oO.
o
050 u- D --
0
0.25 050 075 1.00 1.25 150 1.75 200
o"
.. - 00
125 125
I
.0 i..O.O1 in
o O02m
o 003m
100 100
b..lrr
;AVZ5 pA Y29
f
075
OSO
TE 075
050
o
0.2
o
0.25 OSO 0.75 100 1,25 150
-wYi
t75 2.03
Fig. 3.3. Added massand dampingcocflicicnt in heaving Fig. 3.5. Added mass and damping coefficient in heaving,
LOO
500
0
pA Su
pA
.0
300
o
200
1.00-----
2.
tOy..l
0 2.0
.2.O00n
r
0.0o pA 72g
7
o0
---00 000
s
i',:
05
o
2 5 .61012 S l ..
i lb o
wvi 1.75 200
wni-
Fig. 4.1. Added mass and damping coefficient in swaying
Fig. 4.3. Added mass and damping coefficient in swaying
0
01
pAS 1
pAD
1
05
r:
2
00
_'.uuu000000 060
.50
irr 1.75 200
5.0
010
?tII
pAD 12g
s
-070 000 0 O03m
0
0 000000 o
-0 2 00000
030 D0oa:_ ....
DOD
-0
02 .00510 .515 rr .52 01.0
flc nn
- --wvirr
175 200
2.
H
h
-------.,.-
, -'
00 p
I0o5_-__ -
-6
o o
050
ThD
00
00000 O
---.- -
o
O.lOo 07S
0 25 0.50 0.75 1 00 1.25 1.50
-uvifW
1.75 2.00 0.25 OSO 1.00 1.25 1.50
w
rr 1.75 2.(
1.25
0. BIT 2
O
1,00 O
00 o
. B
ai
bMfr 0'
o
0AV29 . . O pAB
o
07S
050 T
000
-L
L.1
o
025 0.05
001
:
O
pAB
.tL o
pAB
010
-OEs t
00 0 -
- 020 005
0.25 0.1OQ
0.25 OIS
OSO 1.00 1.25 1.50
-w rr 175 2.1
020
0 25 050 075 loo 125 150 175 200
irr
_u_ Yi
_p-.
o
oo
-z
_I
pAD
L
o
_o 00000 0.05
- wyirr
175 2.00 075 1.25 150
0.25 0.50 1.00 175 21
-wYk
Fig. 4.6. Coupling coefflcients of sway;into roll Fig. 4.8. Coupling coefficientsof sway into roll.
268
O y.O,Olm
O O.O2m
O 003m
*
t
0
0 Dy
025 flcO 075 flfl ,n ,,
O i..0.OIm
O 002m
D 003m
1.00
L
0
075 _____ . .
PIO
0 025 OcA 075 inn ic in i,.
-wyli.
Fig. 5.1. Wave amplitude ratio in swaying and heaving
:' H
50
wv
O z..O.OIm
O 002m
O 003m
y.
ESO
I00
250
1.
o
0
J,?
025
O Z..00lm
: :
050 070 1.00 1.25 1.50
wY
O
1.75 2l
1.25
0
0 z.. 001 m
O
0 25
002m
0.50 075 1.00 1.25 1.50
irr
yi.
1.75 2.0
1.00
D 003m
I ;.
025
025
o
o
.50 I. 2.00 1.50 1.75 2.00
irr
wY-i.
Fig. 5.4. Wave amplitude ratio in swaying and heaving Fig. 5.6. Wave amplitude ratio in swaying and heaving
2.
O y.001m O y.001m
002m
________. O
O
D 003m
I I
k
2.00
ISO
l.00
O
050 050
O.
o
O
1.50 b
irr 1.75 2.00 o 0.50 075 1.00 1.25 1.50
Irr I iS 20
Ith
yi
o
,.,
12
O z..0.OIm
o o o
002m
Io 1.00
O
D 003m / o
o
/ o
Z. /0
i::
0
O
o
0.25
D
o.
o
o
rr 075 1.00 125
- tir
1.50 1.7$ 2.00
yi.
Fig. 5.5. Wavc amplitude ratio n swaying and heaving Fig. 5.7. Wave amplitude ratio in swaying and heaving
LIU
02
V
02
0.1
-C)
o,. * $,. 0.05 o
J
o 0.10
o a 0.20
o *
o
t::: 1i
o
o o 005
V
oe
005
D
pou o
o 75 1.25 ISO
__wIr
1.011 1.75 200
a - '2g
-010
1.2 1.50 1.75 2.00 02
O , 0,05
o 0.10
D 0.20
.- .-.- o
o
c
0.050 -- -
D 00
2g o O
0.025 i i: o 000 O
u 0*0
oc 70o000000000
O
O
0(OflO00C00O 005 -D -O20 O
o
0.025e o
025 050 0.75 1.00 1.25 t75
-wv
1.50 2 115 1.50
- " vi
rr 1.75 205
Fig. 6.1. Added mass momcnt of inertia and damping Fig. 6.3. Added mass moment of inertia and damping
*.uciiiici,, iii gull coefficient in roll
-e'-,
o
ut
pA
0.1
o
_o
,.
0
D
O.OS
0.10
0;20
ut
pA
0.1 M
Ii
O
O
i: o i:
o
o O
o V
COO 0.00
0000O RO
005
-0.10
0O D.
C. ODD000000IIUE o
-0.60
.111 00 1.25 150 175 200
o
o0OO'0000000d000c o c o
005 Q...______.
0.1 a
b,.
pA
0,10
pA ij
300 o
O
o
o
0.15
i:: D
0o.Qo0C2.&o
o o,. o.o
o 0 0.10
220e o o 0.20
025 0.50 075 1.00 1.25 1.50 1.75 2.01
OhO
urn, u Q - 1.25 1.50 1.75 200
-w
Fig. 6.2. Coupling coefficients of roll into sway Fig. 6.4. Coupling coefficients of roll into sway
271
0125 0.525
0,100 0.100
a..
i
0025
i::
0.025
o
;-t_
o.
D
or o
n
o
o
o 0
o 050 05 1,00 1.25 1.50 1.75 2.00 0 025 0.50 0.75 1.00 1.25 5.50 175 2 00
w
Ir-
- wi
0125 0.525
o s.. 0.05
o 0.10 m Io s.. 0.05
o ,020 0.0; 020
o too 0100
C
b...lrr
0075 0
o
t::: 0050 o
0025 0025
:
o
L0JjJr88888 8 -'
025 050 075 1.00 1.25 1.50 175 2.00 0.25 050 075 1.00 1.25
rr
1.50
-wvi
1.75 2.
Fig. 6.5: Added mass moment of inertia and damping Fig. 6.7. Added mass moment of inertia and damping
coefflcient in roll coefficient in roll
o 0.55
-005
.!zi
0.io
'fl
O
-- ---
o -010
A9 pAS
-010
.-..' D 0.20
:11 i
o,
ooO
.
i:
o
D
o ?; !i-
-02 0. o-D
00
-010
0--0 025 050 0.75 1.00 1.25 1.50 175 2.00 0.50 0.75 1.00 5Th ISO 1.15 200
lr-
_wvi
ois
o .. 0:05
o 0.10 ; 0.20
005 0.10
oo0 t: .
-
._.it:: - __o__
00_00000
-02 005
025 050 075 1.00 525 1.50 1.75 200 0.50 0.75 1.00 1.25 150 1.75 .. 200
- WY3I
Fig. 6.6. Coupling coefficients of roll into sway Fig. 6.8. Coupling coefficients of roll into sway
2/2
0.150 0.50
0 125
-- -.1
0.'0
0i 2 C,
\ BI,
-
i
o
i::
o
- t::
*
0050
o
OOO
D 0.10
o
o
0.025 D
D
o,.o.os o
0.25 050
r- 0.75 loO
'-H--..-
72
'--- -.
,bo,tO'1 0.10
0.20
o
0.25 0.50 0.75 1.00 1.25 1. 0 1.75 2.00
D.
0.100
F fo ..0.0S 0.'0
o 0.10 2C
Lo 0.20 BI,
0.
t
_2 -. _,
D
D
D
,/_
o-.---
.
O
o
0.20
o
0.25
800000
0.50 075 1.00 1.25
0 O 0.10
- wvii-
Fig. 6.9. Added mass moment of inertia and damping Fig. 7. Wave amplitude ratio in rolling
coefficient in roll
0.15
-- 0-q 050
w'Q:W
0.10
88 - TI
-Sit
pAD
o
o &00
0.3
.---
D
005 o 0.2
o
o u 0.l
o
O
005 o
n
-'r
10 12 16
B
-0.10
0.50 075 1.00 1.25 150 1.15 2.O0
0.150
7
OIS
010
pAD 12g
005 1.00
/AP
0.050 -
-005 0.025
1.50 050
010 o
025 050 .75 1.00 25 .50 12 iL 16
if
-wyi. --r B
Fig. 6.10. Coupling coefficients of roll into sway Fig. 8. Added mass moment of inertia and damping
coefficient in roll vs. B/T-ratio
273
250
200
Y. Y.
eOAk pgAI
t.;: t.
050
O 0
050 075 1.00 1.25 1.50 1.75 2.00 .0 025 050 0.75 tOO 125 1.50 1.75 2.0 o
-
075
- Newman
-- Motora
.1n
1.00
0.75
-- - Newman
Motora
025 025
o
0 025 050 075 100 125 150 175 0 025 050 0.75 .1.00 - 1.25 1.50 175
- wvnr
O 0 0.2 03 0.4 0.5 06 0.7
I
0,9.
I
0.9
I
1.0 lt 0
I
0.1
I
02
I
03
I I
os
I
06
I
07
I
08
I I
.1.0
i
Fig. 9.!. Horizontal and vertical wave exciting forces Fig. 9.3. Horizontal.and vertical wave exciting forces
.
200 200
Y. Y.
p9Ak
-
pgAk L .
150
--9-.3-----.
0 025 0.50' 075 1.00- 1.25 1.50 I.7S 2.0 0.25 0.75
O 0.50 1.00 1.25 1.50 1.75 20 o
2
.5.00
- -
-- Newman
-- Motora
0.75
ii
z.
0.50
025
0.25
-
1
O
050
-wv
0.25 0.75 1.00 1.25 100 1.75 o 0.25 OSO 0.75. 1.00 1.25 1.50 1,75
j02 -
j_
0.3'
I. I j_ I I I
tI, I _I_ I-L I I I I I I
.if
.1 .5 0.6 1.0
.1 U. U.J .4 0.5 06 0.7 0,9 09 1.0 1.1
Fig. 9.2. l-lorizotal and vcrtical waveexcitihg forces Fig. 9.4. Horizontal and vertical wave exciting forces
ff
274
2.0 I
IIu
Ye
1.5 t J
pAk
-
- Newni%
O meuSur.d
ebmt G.
o
t:
00 -A- - o
0 0.25 0.50 .75 1.00 1.25 9.50 9.75
I' -
trr
0 02S 0.50 0.75 1.00 1.25 1.50 1.75 2.0: n n.I n,I --
u
I.
.. -------
I
tj uo
I
U.B 0.7
I
OB
I
0.9
-I
9.0 1.1
i
,yi
I.
Fig. 10.3. Wave exciting moment
Newman
-- Motora
o
z.
Poe C.
05
2 1
':
c.
02
o
0.25 0.50 075 1.00 1.25 1.50 .75
'w
_-_wy.
I J I I I :I_ I- I.- I
U
2
t: o 'i
0
0.25
00
0.50
00
0.75
9
9.00 1.25 .1.50 1.75 K,
3. V.
_wyii kC
' O
Newman
I I I .12
0' nl n
J
I I I J I
1 I 1.
1.0
Z\NiN.wm00 05
o
0 025 0.50 0.75 1.00 1. 5 9.50 I7S
0: 1.0
i I
1.1
1U rj1
90 90
Is
t Q
o
'..
--O Hotoro
80. J.
w
O C
L _.....-----e--
00
o
L. o
-- -------& go. .--- -----4- . o-___,L.o
0 -o 80o0 O
o
O 0.25 0.50 ois loo .1.25 1.50 0.25 0.50 0.75 100 1.25 1.50 1.75 2.0
-. wIn
wyi. lTg-
_1 _l_ I I I I
.7 08. 09 1. Fig. 11.2. Phase angles of the exciting forces and moment
go.
o --
I
O
I
Motora
L
.
- .0.O_1
o 8 . .
O
e.
90
_f:b_;_._O_
180!
0 0.25 0.50 o 0.75 1.00 .1.25 1.50 1.75
w
-'w
I I I I I I -I
0. 0.8 0.7 .8 1. 1.
au
00 0t0
leo.
.2 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.0
-e -- Motora
t,
u:.
__S_o,- -- 0
jO
.-
o
__!oi o
-1__
0
-r
0 0.25 0.'0 0.75 1.00 125 1.50 1.75
.wIn
I I I I I I I i
a'. 0.2 0.3 01. 0.5 0.5 0.7 0.8 1.o 1.1
Roll amplitude
w Frequency of motion or wave frequency